Silence Speaks Volumes: Re-weighting Techniques for Under-Represented Users in Fake News Detection

TL;DR

This paper addresses the bias towards active users in fake news detection models by proposing re-weighting techniques to incorporate silent users, aiming to improve detection accuracy.

Contribution

It introduces re-weighting methods to balance user influence, emphasizing silent users' cues to enhance fake news detection models.

Findings

Re-weighting improves fake news detection accuracy.

Silent users' cues contribute significantly to model performance.

Addresses participation inequality bias in social media analysis.

Abstract

Social media platforms provide a rich environment for analyzing user behavior. Recently, deep learning-based methods have been a mainstream approach for social media analysis models involving complex patterns. However, these methods are susceptible to biases in the training data, such as participation inequality. Basically, a mere 1% of users generate the majority of the content on social networking sites, while the remaining users, though engaged to varying degrees, tend to be less active in content creation and largely silent. These silent users consume and listen to information that is propagated on the platform. However, their voice, attitude, and interests are not reflected in the online content, making the decision of the current methods predisposed towards the opinion of the active users. So models can mistake the loudest users for the majority. We propose to leverage re-weighting techniques to make the silent majority heard, and in turn, investigate whether the cues from these users can improve the performance of the current models for the downstream task of fake news detection.